keras-neural-graph-fingerprint
keras-neural-graph-fingerprint copied to clipboard
keiserlab
Keras implementation of Neural Graph Fingerprints as proposed by Duvenaud et al., 2015
There are quite a few issues when using your code with Keras 2. For example, the customised Keras layers does not work and gives the following issue: File "C:\Users\User\Desktop\Python\SNN_cont_only\own_package\NGF\layers.py", line...
This version is compatible with both Keras 2 and Tensorflow backend. All the examples run as expected. The `filter_func_args` function inside utils still needs to adapt the Keras 2 style...
I replaced the Theano backend with Tensorflow and commented out the Theano from NGF/layers.py. The code works fine for example #1. Also, I am going to make it compatible with...
The only theano dependency is due to the modification of `temporal_padding` in [NGF/layers.py]( https://github.com/keiserlab/keras-neural-graph-fingerprint/blob/92a5807dc31630f4fee338ef6c1b01a51786e472/NGF/layers.py#L16) See if the desired modification is supported in keras 2 (after fixing #10) and otherwise implement...
As @8li pointed out, the current code is incompatible with keras 2 due to the `layer_from_config` in [NGF/layers.py](https://github.com/keiserlab/keras-neural-graph-fingerprint/blob/92a5807/NGF/layers.py). Investigate what other issues arise when using keras 2 and see if...
The `SparseTensor`s implemented in [NGF/sparse.py](NGF/sparse.py) are very well suited for the molecule data structures/distributions used in these repo. An advantage of the SparseTensors tensors is that they do not require...
In the [neural graph fingerprint paper](https://arxiv.org/abs/1509.09292), and [implementation](https://github.com/HIPS/neural-fingerprint), the summed features for a node are a concatenation of the summed atom features of the neighbours, and the summed bond features...